Presented are systems and methods to automatically find efficient parallelization strategies for deep neural networks (DNNs). A computation graph comprising an efficiently ordered sequence of vertices aids in computing the best parallelizing strategy in a relatively short time. Effectiveness of the parallelization strategies is evaluated on various DNNs, and the performance of the strategies proposed by various embodiments is compared against data parallelism, expert-designed strategies, and other state-of-the-art approaches. Experimental results demonstrate that the proposed strategies outperform a baseline data parallelism strategy and achieve better performance than expert-designed strategies and state-of-the-art approaches.

Provided are an apparatus and a method for providing screen setting data of a plurality of devices. The apparatus includes at least: a communication unit that transmits and receives data; a storage unit that stores the data; and a controller operatively connected to the communication unit and the storage unit. The controller receives screen configuration data determined by a manufacturer in association with a specific device selected by the manufacturer of a first user device from the first user device, stores screen setting data based on the received screen configuration data and screen characteristic data for a predetermined standard screen of the specific device in the storage unit, receives a screen setting request for the specific device from a second user device, and transmits screen setting data stored so as to correspond to the specific device in accordance with the screen setting request to the specific device.

Embodiments of the present disclosure provide a method and apparatus for processing a media type in representational state transfer (REST) software architecture, comprising: extracting, in a request from a client, a first media type for the request, the request including a message sent to a server; in response to determining that a first media type processor supporting the first media type exists, converting, by the first media type processor, the message into an instance of a first data model; and sending the instance of the first data model to the server. Embodiments of the present disclosure can add support to a new media type without a need of changing the existing data model, thereby enhancing development efficiency of web applications.

Embodiments of the present disclosure provide a method and apparatus for processing a media type in representational state transfer (REST) software architecture, comprising: extracting, in a request from a client, a first media type for the request, the request including a message sent to a server; in response to determining that a first media type processor supporting the first media type exists, converting, by the first media type processor, the message into an instance of a first data model; and sending the instance of the first data model to the server. Embodiments of the present disclosure can add support to a new media type without a need of changing the existing data model, thereby enhancing development efficiency of web applications.

In general, certain embodiments of the present disclosure provide methods and systems for enabling a reproducible processing of machine learning models and scalable deployment on a distributed network. The method comprises building a machine learning model; training the machine learning model to produce a plurality of versions of the machine learning model; tracking the plurality of versions of the machine learning model to produce a change facilitator tool; sharing the change facilitator tool to one or more devices such that each device can reproduce the plurality of versions of the machine learning model; and generating a deployable version of the machine learning model through repeated training.

In general, certain embodiments of the present disclosure provide methods and systems for enabling a reproducible processing of machine learning models and scalable deployment on a distributed network. The method comprises building a machine learning model; training the machine learning model to produce a plurality of versions of the machine learning model; tracking the plurality of versions of the machine learning model to produce a change facilitator tool; sharing the change facilitator tool to one or more devices such that each device can reproduce the plurality of versions of the machine learning model; and generating a deployable version of the machine learning model through repeated training.

A computer-implemented method includes receiving, by a computing device, an event notification; determining, by the computing device, whether to immediately execute a pipeline including a commit associated with the event notification based on historical trends of commits entering the pipeline at a similar time period as a current time; and immediately executing or delaying the execution of the pipeline, by the computing device, based on the determining whether to immediately execute the pipeline.

A system and method for defending an application configured to invoke anonymous functions. The method includes analyzing the application to determine at least one branch of the application, wherein each branch is an instruction that deviates from a default behavior of the application; identifying, based on the at least one branch of the application and at least one first anonymous function, at least one potential threat branch, each potential threat branch including a call to one of the at least one first anonymous function; and rewiring at least one first function call of the application to create a secured instance of the application, wherein each of the at least one first function call is to one of the at least one first anonymous function prior to rewiring.

A system and method for defending an application configured to invoke anonymous functions. The method includes analyzing the application to determine at least one branch of the application, wherein each branch is an instruction that deviates from a default behavior of the application; identifying, based on the at least one branch of the application and at least one first anonymous function, at least one potential threat branch, each potential threat branch including a call to one of the at least one first anonymous function; and rewiring at least one first function call of the application to create a secured instance of the application, wherein each of the at least one first function call is to one of the at least one first anonymous function prior to rewiring.

An adaptive content platform includes one or more content-enabled, dependent applications, each of which includes a user interface and business logic. A services layer, which is interfaced with the dependent applications and a software infrastructure, provides one or more services that are usable by the dependent applications.